netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Vladimir Oltean <olteanv@gmail.com>
To: DENG Qingfang <dqfext@gmail.com>
Cc: "Sean Wang" <sean.wang@mediatek.com>,
	"Landen Chao" <Landen.Chao@mediatek.com>,
	"Andrew Lunn" <andrew@lunn.ch>,
	"Vivien Didelot" <vivien.didelot@gmail.com>,
	"Florian Fainelli" <f.fainelli@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	"Jakub Kicinski" <kuba@kernel.org>,
	"Matthias Brugger" <matthias.bgg@gmail.com>,
	netdev@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org,
	"Eric Woudstra" <ericwouds@gmail.com>,
	"René van Dorst" <opensource@vdorst.com>,
	"Frank Wunderlich" <frank-w@public-files.de>
Subject: Re: [RFC net-next v2 2/4] net: dsa: mt7530: use independent VLAN learning on VLAN-unaware bridges
Date: Mon, 2 Aug 2021 16:36:13 +0300	[thread overview]
Message-ID: <20210802133613.aeyduvt6qvy2ei7r@skbuf> (raw)
In-Reply-To: <20210731191023.1329446-3-dqfext@gmail.com>

On Sun, Aug 01, 2021 at 03:10:20AM +0800, DENG Qingfang wrote:
> Consider the following bridge configuration, where bond0 is not
> offloaded:
> 
>          +-- br0 --+
>         / /   |     \
>        / /    |      \
>       /  |    |     bond0
>      /   |    |     /   \
>    swp0 swp1 swp2 swp3 swp4
>      .        .       .
>      .        .       .
>      A        B       C
> 
> Ideally, when the switch receives a packet from swp3 or swp4, it should
> forward the packet to the CPU, according to the port matrix and unknown
> unicast flood settings.
> 
> But packet loss will happen if the destination address is at one of the
> offloaded ports (swp0~2). For example, when client C sends a packet to
> A, the FDB lookup will indicate that it should be forwarded to swp0, but
> the port matrix of swp3 and swp4 is configured to only allow the CPU to
> be its destination, so it is dropped.
> 
> However, this issue does not happen if the bridge is VLAN-aware. That is
> because VLAN-aware bridges use independent VLAN learning, i.e. use VID
> for FDB lookup, on offloaded ports. As swp3 and swp4 are not offloaded,
> shared VLAN learning with default filter ID of 0 is used instead. So the
> lookup for A with filter ID 0 never hits and the packet can be forwarded
> to the CPU.
> 
> In the current code, only two combinations were used to toggle user
> ports' VLAN awareness: one is PCR.PORT_VLAN set to port matrix mode with
> PVC.VLAN_ATTR set to transparent port, the other is PCR.PORT_VLAN set to
> security mode with PVC.VLAN_ATTR set to user port.
> 
> It turns out that only PVC.VLAN_ATTR contributes to VLAN awareness, and
> port matrix mode just skips the VLAN table lookup. The reference manual
> is somehow misleading when describing PORT_VLAN modes. It states that
> PORT_MEM (VLAN port member) is used for destination if the VLAN table
> lookup hits, but actually **PORT_MEM & PORT_MATRIX** (bitwise AND of
> VLAN port member and port matrix) is used instead, which means we can
> have two or more separate VLAN-aware bridges with the same PVID and
> traffic won't leak between them.

"traffic won't leak" is only half the story. It won't leak, but when
multiple VLAN-unaware bridges share the same FID, they are still subject
to shortcircuit attempts if the same MAC address is present in the FDB
of both bridges. This patch doesn't solve that, maybe it is worth adding
a big comment in the code itself that clarifies that FDBs between
multiple VLAN-unaware bridges, as well as between multiple VLAN-aware
bridges, are not isolated per se, only that standalone ports are placed
under a different FID compared to bridges of whatever kind.

> 
> Therefore, to solve this, enable independent VLAN learning with PVID 0
> on VLAN-unaware bridges, by setting their PCR.PORT_VLAN to fallback
> mode, while leaving standalone ports in port matrix mode. The CPU port
> is always set to fallback mode to serve those bridges.
> 
> During testing, it is found that FDB lookup with filter ID of 0 will
> also hit entries with VID 0 even with independent VLAN learning. To
> avoid that, install all VLANs with filter ID of 1.
> 
> Signed-off-by: DENG Qingfang <dqfext@gmail.com>
> ---
>  drivers/net/dsa/mt7530.c | 70 ++++++++++++++++++++++++++++------------
>  drivers/net/dsa/mt7530.h |  4 ++-
>  2 files changed, 53 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/net/dsa/mt7530.c b/drivers/net/dsa/mt7530.c
> index 7e7e0a35e351..3876e265f844 100644
> --- a/drivers/net/dsa/mt7530.c
> +++ b/drivers/net/dsa/mt7530.c
> @@ -1021,6 +1021,10 @@ mt753x_cpu_port_enable(struct dsa_switch *ds, int port)
>  	mt7530_write(priv, MT7530_PCR_P(port),
>  		     PCR_MATRIX(dsa_user_ports(priv->ds)));
>  
> +	/* Set to fallback mode for independent VLAN learning */
> +	mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> +		   MT7530_PORT_FALLBACK_MODE);
> +
>  	return 0;
>  }
>  
> @@ -1229,6 +1233,10 @@ mt7530_port_bridge_join(struct dsa_switch *ds, int port,
>  			   PCR_MATRIX_MASK, PCR_MATRIX(port_bitmap));
>  	priv->ports[port].pm |= PCR_MATRIX(port_bitmap);
>  
> +	/* Set to fallback mode for independent VLAN learning */
> +	mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> +		   MT7530_PORT_FALLBACK_MODE);
> +
>  	mutex_unlock(&priv->reg_mutex);
>  
>  	return 0;
> @@ -1241,16 +1249,21 @@ mt7530_port_set_vlan_unaware(struct dsa_switch *ds, int port)
>  	bool all_user_ports_removed = true;
>  	int i;
>  
> -	/* When a port is removed from the bridge, the port would be set up
> -	 * back to the default as is at initial boot which is a VLAN-unaware
> -	 * port.
> +	/* This is called after .port_bridge_leave when leaving a VLAN-aware
> +	 * bridge. Don't set standalone ports to fallback mode.
>  	 */
> -	mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> -		   MT7530_PORT_MATRIX_MODE);
> +	if (dsa_to_port(ds, port)->bridge_dev)
> +		mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> +			   MT7530_PORT_FALLBACK_MODE);
> +

hmm, okay, the ordering between dsa_broadcast() and
dsa_port_switchdev_unsync_attrs() in dsa_port_bridge_leave() is a bit
awkward for you, but this looks okay.

>  	mt7530_rmw(priv, MT7530_PVC_P(port), VLAN_ATTR_MASK | PVC_EG_TAG_MASK,
>  		   VLAN_ATTR(MT7530_VLAN_TRANSPARENT) |
>  		   PVC_EG_TAG(MT7530_VLAN_EG_CONSISTENT));
>  
> +	/* Set PVID to 0 */
> +	mt7530_rmw(priv, MT7530_PPBV1_P(port), G0_PORT_VID_MASK,
> +		   G0_PORT_VID_DEF);
> +
>  	for (i = 0; i < MT7530_NUM_PORTS; i++) {
>  		if (dsa_is_user_port(ds, i) &&
>  		    dsa_port_is_vlan_filtering(dsa_to_port(ds, i))) {
> @@ -1276,15 +1289,14 @@ mt7530_port_set_vlan_aware(struct dsa_switch *ds, int port)
>  	struct mt7530_priv *priv = ds->priv;
>  
>  	/* Trapped into security mode allows packet forwarding through VLAN
> -	 * table lookup. CPU port is set to fallback mode to let untagged
> -	 * frames pass through.
> +	 * table lookup.
>  	 */
> -	if (dsa_is_cpu_port(ds, port))
> -		mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> -			   MT7530_PORT_FALLBACK_MODE);
> -	else
> +	if (dsa_is_user_port(ds, port)) {
>  		mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
>  			   MT7530_PORT_SECURITY_MODE);
> +		mt7530_rmw(priv, MT7530_PPBV1_P(port), G0_PORT_VID_MASK,
> +			   G0_PORT_VID(priv->ports[port].pvid));
> +	}
>  
>  	/* Set the port as a user port which is to be able to recognize VID
>  	 * from incoming packets before fetching entry within the VLAN table.
> @@ -1329,6 +1341,13 @@ mt7530_port_bridge_leave(struct dsa_switch *ds, int port,
>  			   PCR_MATRIX(BIT(MT7530_CPU_PORT)));
>  	priv->ports[port].pm = PCR_MATRIX(BIT(MT7530_CPU_PORT));
>  
> +	/* When a port is removed from the bridge, the port would be set up
> +	 * back to the default as is at initial boot which is a VLAN-unaware
> +	 * port.
> +	 */
> +	mt7530_rmw(priv, MT7530_PCR_P(port), PCR_PORT_VLAN_MASK,
> +		   MT7530_PORT_MATRIX_MODE);
> +
>  	mutex_unlock(&priv->reg_mutex);
>  }
>  
> @@ -1511,7 +1530,7 @@ mt7530_hw_vlan_add(struct mt7530_priv *priv,
>  	/* Validate the entry with independent learning, create egress tag per
>  	 * VLAN and joining the port as one of the port members.
>  	 */
> -	val = IVL_MAC | VTAG_EN | PORT_MEM(new_members) | VLAN_VALID;
> +	val = IVL_MAC | VTAG_EN | PORT_MEM(new_members) | FID(1) | VLAN_VALID;
>  	mt7530_write(priv, MT7530_VAWD1, val);
>  
>  	/* Decide whether adding tag or not for those outgoing packets from the
> @@ -1601,9 +1620,13 @@ mt7530_port_vlan_add(struct dsa_switch *ds, int port,
>  	mt7530_hw_vlan_update(priv, vlan->vid, &new_entry, mt7530_hw_vlan_add);
>  
>  	if (pvid) {
> -		mt7530_rmw(priv, MT7530_PPBV1_P(port), G0_PORT_VID_MASK,
> -			   G0_PORT_VID(vlan->vid));
>  		priv->ports[port].pvid = vlan->vid;
> +
> +		/* Only configure PVID if VLAN filtering is enabled */
> +		if (dsa_port_is_vlan_filtering(dsa_to_port(ds, port)))
> +			mt7530_rmw(priv, MT7530_PPBV1_P(port),
> +				   G0_PORT_VID_MASK,
> +				   G0_PORT_VID(vlan->vid));
>  	}
>  
>  	mutex_unlock(&priv->reg_mutex);
> @@ -1617,11 +1640,9 @@ mt7530_port_vlan_del(struct dsa_switch *ds, int port,
>  {
>  	struct mt7530_hw_vlan_entry target_entry;
>  	struct mt7530_priv *priv = ds->priv;
> -	u16 pvid;
>  
>  	mutex_lock(&priv->reg_mutex);
>  
> -	pvid = priv->ports[port].pvid;
>  	mt7530_hw_vlan_entry_init(&target_entry, port, 0);
>  	mt7530_hw_vlan_update(priv, vlan->vid, &target_entry,
>  			      mt7530_hw_vlan_del);
> @@ -1629,11 +1650,12 @@ mt7530_port_vlan_del(struct dsa_switch *ds, int port,
>  	/* PVID is being restored to the default whenever the PVID port
>  	 * is being removed from the VLAN.
>  	 */
> -	if (pvid == vlan->vid)
> -		pvid = G0_PORT_VID_DEF;
> +	if (priv->ports[port].pvid == vlan->vid) {
> +		priv->ports[port].pvid = G0_PORT_VID_DEF;
> +		mt7530_rmw(priv, MT7530_PPBV1_P(port), G0_PORT_VID_MASK,
> +			   G0_PORT_VID_DEF);
> +	}
>  
> -	mt7530_rmw(priv, MT7530_PPBV1_P(port), G0_PORT_VID_MASK, pvid);
> -	priv->ports[port].pvid = pvid;
>  
>  	mutex_unlock(&priv->reg_mutex);
>  
> @@ -2134,6 +2156,10 @@ mt7530_setup(struct dsa_switch *ds)
>  				return ret;
>  		} else {
>  			mt7530_port_disable(ds, i);
> +
> +			/* Set default PVID to 0 on all user ports */
> +			mt7530_rmw(priv, MT7530_PPBV1_P(i), G0_PORT_VID_MASK,
> +				   G0_PORT_VID_DEF);
>  		}
>  		/* Enable consistent egress tag */
>  		mt7530_rmw(priv, MT7530_PVC_P(i), PVC_EG_TAG_MASK,
> @@ -2301,6 +2327,10 @@ mt7531_setup(struct dsa_switch *ds)
>  				return ret;
>  		} else {
>  			mt7530_port_disable(ds, i);
> +
> +			/* Set default PVID to 0 on all user ports */
> +			mt7530_rmw(priv, MT7530_PPBV1_P(i), G0_PORT_VID_MASK,
> +				   G0_PORT_VID_DEF);
>  		}
>  
>  		/* Enable consistent egress tag */
> diff --git a/drivers/net/dsa/mt7530.h b/drivers/net/dsa/mt7530.h
> index b19b389ff10a..a308886fdebc 100644
> --- a/drivers/net/dsa/mt7530.h
> +++ b/drivers/net/dsa/mt7530.h
> @@ -148,6 +148,8 @@ enum mt7530_vlan_cmd {
>  #define  VTAG_EN			BIT(28)
>  /* VLAN Member Control */
>  #define  PORT_MEM(x)			(((x) & 0xff) << 16)
> +/* Filter ID */
> +#define  FID(x)				(((x) & 0x7) << 1)
>  /* VLAN Entry Valid */
>  #define  VLAN_VALID			BIT(0)
>  #define  PORT_MEM_SHFT			16
> @@ -247,7 +249,7 @@ enum mt7530_vlan_port_attr {
>  #define MT7530_PPBV1_P(x)		(0x2014 + ((x) * 0x100))
>  #define  G0_PORT_VID(x)			(((x) & 0xfff) << 0)
>  #define  G0_PORT_VID_MASK		G0_PORT_VID(0xfff)
> -#define  G0_PORT_VID_DEF		G0_PORT_VID(1)
> +#define  G0_PORT_VID_DEF		G0_PORT_VID(0)
>  
>  /* Register for port MAC control register */
>  #define MT7530_PMCR_P(x)		(0x3000 + ((x) * 0x100))
> -- 
> 2.25.1
> 

Reviewed-by: Vladimir Oltean <olteanv@gmail.com>

  reply	other threads:[~2021-08-02 13:36 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-31 19:10 [RFC net-next v2 0/4] mt7530 software fallback bridging fix DENG Qingfang
2021-07-31 19:10 ` [RFC net-next v2 1/4] net: dsa: mt7530: enable assisted learning on CPU port DENG Qingfang
2021-08-02 13:07   ` Vladimir Oltean
2021-07-31 19:10 ` [RFC net-next v2 2/4] net: dsa: mt7530: use independent VLAN learning on VLAN-unaware bridges DENG Qingfang
2021-08-02 13:36   ` Vladimir Oltean [this message]
2021-07-31 19:10 ` [RFC net-next v2 3/4] net: dsa: mt7530: set STP state also on filter ID 1 DENG Qingfang
2021-08-02 13:43   ` Vladimir Oltean
2021-08-02 15:31     ` DENG Qingfang
2021-08-02 15:42       ` Vladimir Oltean
2021-08-02 15:58         ` DENG Qingfang
2021-08-02 21:00           ` Vladimir Oltean
2021-08-03  8:23             ` DENG Qingfang
2021-08-03  8:48               ` Vladimir Oltean
2021-07-31 19:10 ` [RFC net-next v2 4/4] Revert "mt7530 mt7530_fdb_write only set ivl bit vid larger than 1" DENG Qingfang
2021-08-02 13:44   ` Vladimir Oltean
2021-08-02 15:48     ` DENG Qingfang
2021-08-02 20:59       ` Vladimir Oltean

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210802133613.aeyduvt6qvy2ei7r@skbuf \
    --to=olteanv@gmail.com \
    --cc=Landen.Chao@mediatek.com \
    --cc=andrew@lunn.ch \
    --cc=davem@davemloft.net \
    --cc=dqfext@gmail.com \
    --cc=ericwouds@gmail.com \
    --cc=f.fainelli@gmail.com \
    --cc=frank-w@public-files.de \
    --cc=kuba@kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=matthias.bgg@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=opensource@vdorst.com \
    --cc=sean.wang@mediatek.com \
    --cc=vivien.didelot@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).